Resolving host and species boundaries for perithecia-producing nectriaceous fungi across the central Appalachian Mountains

The Nectriaceae contains numerous canker pathogens. Due to scarcity of ascomata on many hosts, comprehensive surveys are lacking. Here we characterize the diversity of perithecia-producing nectriaceous fungi across the central Appalachians. Ten species from twelve hosts were recovered including a novel Corinectria sp. from Picea rubens. Neonectria ditissima and N. faginata were most abundant and associated with Fagus grandifolia with beech bark disease (BBD). N. ditissima was also recovered from additional cankered hardwoods, including previously unreported Acer spicatum, Ilex mucronata, and Sorbus americana. Cross-pathogenicity inoculations of N. ditissima confirmed susceptibility of Acer and Betula spp. Neonectria magnoliae was recovered from cankered Liriodendron tulipifera and Magnolia fraseri and pathogenicity on L. tulipifera was confirmed. Fusarium babinda was consistently recovered from beech with BBD, although its role remains unclear. This survey provides a contemporary snapshot of Nectriaceae diversity across the Appalachian Mountains. The following nomenclatural changes are proposed: Neonectria magnoliae comb. nov.


Introduction
Members of the Nectriaceae occupy diverse ecological niches from mycoparasites to phytopathogens, with numerous genera and species implicated in causing annual and perennial cankers on diverse woody plant hosts with varying degrees of host specificity. One such canker disease, beech bark disease (BBD), is a disease complex occurring across the range of American beech (Fagus grandifolia American beech or N. coccinea ([Pers.] Rossman and Samuels) (Houston, 1994b;Thomsen et al., 1949). In North America, BBD was first observed in Halifax, Nova Scotia, Canada around 1890 and presently continues to spread throughout the continuous range of American beech (Hewitt, 1914;Erhlich, 1934;. Other members of Nectriaceae have occasionally been associated with BBD, including underestimating diversity (for example see Fig 2 in Kasson and Livingston, 2009).
Likewise, the close relationships among Neonectria spp., coupled with the lack of sequence data in repositories such as NCBI, has also presented species identification challenges, especially for rarer species, which further complicates identification. For example, our recent phylogenetic analyses of mating type gene sequences for Nectria magnoliae from Liriodendron tulipifera L. revealed this species formed a genealogically exclusive clade among other Neonectria species that was closely allied with N. faginata (Stauder et al., 2020). This supports previous findings by Gräfenhan and colleagues (2011), who showed but did not discuss that N. ditissima and N. magnoliae were genealogically exclusive. However, earlier work by Castlebury et al. (2006) concluded N. magnoliae was a synonym of N. ditissima. These findings emphasize the need for enhanced surveys to recover and phylogenetically resolve cryptic and understudied members of Nectriaceae as well as investigate the potential for cryptic reservoirs of known species including the possibility of N. faginata occurring on a non-beech host.
This study sought to provide a greater understanding of the ecology and genetic relationships among N. ditissima, N. faginata, and allied fungi. The first objective of this study was to survey perithecia-producing members of Nectriaceae in American beech stands to identify possible native reservoirs of N. faginata. This is important given the uncertainty surrounding the origin of N. faginata and the potential for previous misidentifications of nectriaceous fungi recovered from non-beech hosts. The second objective was to resolve phylogenetic relationships among described and possibly undescribed members of Nectriaceae recovered from forests across the central Appalachian Mountains. This was critical given both the significant undersampling of nectriaceous fungi in these regions, the resulting the lack of sequence data, and/or contradictory evidence regarding certain known members of Nectriaceae, such as Nectria magnoliae. A third objective was to test host specificity of N. ditissima, N. faginata, and N. magnoliae strains isolated from American beech trees and tulip poplar.
Together these aims sought to provide contemporary insights into the true diversity, phylogenetic relationships, and ecological niches of perithecia-producing nectriaceous fungi across the central Appalachian Mountains.

Site Selection
Sampling locations were selected based on previous reports of cankers and/or perithecia production on woody hosts, accessibility, and ability to secure sampling permits. In total, 13 beech bark disease (BBD) and eight non-BBD sampling locations were selected across WV, MD, VA, PA, TN and NC ( Fig. 1; Supplemental Table 1).
Four of the eight non-BBD sites served for sampling nectriaceous fungi from non-beech hosts, while the other four served as sites where beech tissue samples were collected from stands with no history of BBD. No defined sampling areas (i.e. plots) were established, but instead, trees were sampled opportunistically until sampling needs were met based on the sampling procedure described below. Sites GK, MM, and SM had five, two, and four geographically separated discrete sampling areas (Supplemental Table 1). All other sites are represented by a single sampling area.
Three WV sites (BM, GK, FR) were originally confirmed as having BBD at least 35 years prior to the initiation of this study , Houston, 1994a. Sites designated BG, DM, and PSF represented more recent BBD outbreaks (~ 15 years) as determined by various natural resource agency surveys in WV and MD. All other sites were determined to have likely become afflicted with BBD between 5 to 25 years prior (Morin et al., 2007). Four of seven sites lacking BBD (NB, PM, AR, UW) were selected for asymptomatic beech sampling based on a lack of previous reports, and all trees were visually confirmed to be absent of BBD-associated signs and symptoms prior to sampling. previously reported are filled. BBD was first discovered in WV at site GK in 1981 Houston, 1983). BBD distribution is based on a county level database (Blackburn, 2018).

Sampling Procedure
At each BBD site, American beech trees symptomatic for BBD and non-beech trees with perennial target cankers were identified and sampled. For approximately five trees of each species per site, a maximum of four bark disks harboring fresh perithecia were excised with a 1-cm diameter steel punch and stored in a microtiter plates.
Additionally, bark tissue samples (~1 mm diameter plugs; ~8 samples/tree; up to 7 beech trees/site) were taken from symptomatic and asymptomatic trees at BBD and non-BBD sites, respectively. All samples were stored on ice until arriving at the laboratory where samples were stored at -20 °C until processing.

Sample Processing
To process each bark disk, up to five perithecia were removed with a sterile scalpel and cleaned by gently pushing each perithecium through sterile agar to remove debris. Each perithecium was squashed in a 1.5 ml microcentrifuge tube containing 1 ml of sterile H 2 O with a micropestle, vortexed for 15 seconds, and then 300 µl of the spore suspension was spread with a cell spreader on glucose-yeast extract agar plates amended with streptomycin sulfate (10 mg/1000 ml) and tetracycline hydrochloride (100 mg/1000 ml) antibiotics (GYE/A). Within 48 hours of plating, five germinating ascospores were subcultured to a new plate, and one isolate was selected for storage at approximately -20 °C on glass filter paper. All micro-sampled bark plugs were surface disinfested by soaking for 14 minutes in a 1:10 commercial bleach-water solution then up to four samples were placed onto each GYE/A agar plate. Resulting fungi of interest were subcultured individually to new plates then stored as previously described.

Species Identification
Recovered isolates were grouped and tentatively identified based on colony and macroconidia morphology and a subset of isolates were selected to be confirmed by DNA sequencing. Genomic DNA was extracted from isolates using a Wizard® kit (Promega, Madison, WI, USA) and suspended in 75 µl Tris-EDTA (TE) buffer (Amresco, Solon, OH, USA). PCR reactions were performed for the fungal barcoding genes internal transcribed spacer region (ITS), translation elongation factor 1-alpha (EF-1α), 28S rDNA (LSU), and β -tubulin (BTUB sequences were used to identify species using BLASTn searches, and the best match in the NCBI database was selected for the isolate's identity.

Phylogenetic Analyses
For all sequences included in our phylogenetic analyses, chromatograms were assessed and clipped using CodonCode Aligner v. 5.1.5. Sequences were then manually corrected for nucleotide misreads by referencing their respective chromatograms. To examine phylogenetic relationships among collected members of Nectriaceae, single-gene and concatenated phylogenetic trees were constructed for all Corinectria, Fusarium, and Neonectria species observed in this survey along with additional reference sequences for selected members of Nectriaceae available from NCBI Genbank (Table 3). The other members of the Nectriaceae recovered either lacked sampling depth or adequate reference sequences for our loci of interest to permit meaningful phylogenetic analysis. As such, only BLASTn searches of individual loci were conducted for these species with representative sequences deposited into NCBI Genbank. For Corinectria, Fusarium, and Neonectria species, each gene was aligned using MAFFT (Katoh and Stanley, 2013) on the Guidance 2.0 server (Landan and Graur, 2008;Sela et al., 2015). Individual resides with Guidance scores <0.5 were masked (Macias et al., 2020). A concatenated sequence was generated from single gene alignments using the web tool FaBox (Villesen, 2007).
For single gene and concatenated sequences, maximum-likelihood analyses were completed using MEGA v10.1.7 (Stecher et al., 2020), and Bayesian inference (BI) analyses were completed using MrBayes v. 3.2.7 (Ronquist et al., 2012). For ML analyses, the best-fit nucleotide substitution model was chosen using Model Test AICc scores in MEGA and 1000 bootstrap replicates were used. For BI analyses, MrBayes selected the best fit nucleotide selection model, but the rate of substitution was selected from the Model test AICc scores. The BI runs were stopped once the standard deviation of split frequencies fell below 0.01. These were then checked for convergence in Tracer v. 1.7.1 (Rambaut et al., 2018). Trees were prepared for publication using FigTree v.

Morphological Measurements
Over the course of this survey, a species molecularly identified as Nectria magnoliae was recovered from tulip poplar (Liriodendron tulipifera L.) and Fraser magnolia (Magnolia fraseri Walter). Based on a recent study in which isolates from both of these hosts were included, this species was shown to form an independent clade with other members of Neonectria (Stauder et al., 2020). A previously published study also supported these relationships (Gräfenhan et al. 2011). Given these phylogenetic relationships, ascospore and conidia measurements were conducted to further characterize this species to permit comparisons with original descriptions provided by Lohman and Watson (1943).
For ascospore measurements, a total of 15 perithecia were processed. Three perithecia were sampled from each of five bark disks representing two geographically separated sites (FN = 3 disks; SH = 2 disks). A single perithecium was extracted from the bark disk using a sterile scalpel and squash-mounted on a slide with lactophenol plus cotton blue mountant. Length and width measurements were collected for 25 ascospores per perithecium. All measurements were taken with a Nikon Eclipse E600 compound microscope (Nikon Instruments, Melville, NY, USA) equipped with a Nikon Digital Sight DS-Ri1 microscope camera and Nikon NIS-Elements BR3.2 imaging software.
Conidia measurements were similarly conducted for both micro-and macroconidia. Here, sporodochial masses were harvested from pure four-to six-week-old cultures using a sterile scalpel and mounted onto slides as described above. Both length and width measurements were taken for 50 microconidia were harvested from two isolates from each of three geographically separated locations (GK, SH, FN).
Samples from SH and FN were collected from L. tulipifera, and samples from GK were collected from M. fraseri. Macroconidia were only found associated with one isolate collected from M. fraseri at GK and measured as described.

Pathogenicity Assays
Field inoculations were conducted to further investigate pathogenicity of N.
magnoliae strains from L. tulipifera as well as N. faginata and N. ditissima strains from beech on birch, striped maple, and beech. For N. magnoliae, a single isolate (NmLt001) recovered from a natural infection on L. tulipifera was selected. One isolate of both N.
ditissima (NdFg002) and N. faginata (NfFg005) recovered from American beech were also included for cross-pathogenicity testing. All study isolates were grown in pure culture on GYE for two-weeks at room temperature. Prior to inoculations, a sterile 1-cm steel punch was used to cut inoculation plugs along the growing edge of the colony.
Negative control plugs were cut from sterile GYE plates.
Six each of tulip poplar (L. tulipifera), black birch (Betula lenta), yellow birch (Betula alleghaniensis Britt.), and striped maple (Acer pensylvanicum) trees, located on WVU University Forest were selected for inoculations. Additionally, five American beech trees were selected at this same location. Each tree received an inoculation with N. ditissima, N. faginata, and N. magnoliae. In addition to these fungal inoculations, a negative control inoculation with a sterile GYE agar plug was also performed on each study tree.
For each inoculations, a sterile 1-cm leather punch was used to excise bark tissue and create an inoculum reservoir. A colonized or sterile agar plug was then placed into the reservoir, and masking tape was applied over the wound to limit inoculum desiccation prior to infection. After six-months, bark tissue was excised from canker margins using a bone-marrow biopsy tool and placed in a 96-well microtiter dish.
To ensure accurate canker measurements, a knife was used to remove bark tissue and reveal any underlying necrosis. Length and width measurements were then taken for each canker resulting from inoculation. Recovery of inoculant was achieved as previously described for bark tissue sample processing. All isolate identities were confirmed morphologically.

Statistical analyses
For pathogenicity measures, a one-way ANOVA was completed to check for differences in canker size and a Tukey-HSD post-hoc test was completed identify significant pairwise differences using the stats v3.6.2 package within R v 3.6.3 statistical software (R Core Team, 2020). All p-values <0.05 were considered significant.

Survey of nectriaceous fungi
Perithecia of putative nectriaceous fungi were sampled from 180 trees across 17 sites in WV, MD, VA, NC, and TN (Table 1). Sampling included 12 tree species with black birch, red spruce, and mountain ash being the most abundantly sampled species besides America beech. A total of 1,605 sampled perithecia yielded nine fungal species spanning six genera. The majority of these samples were collected from American beech trees (n = 1,257 perithecia). The remaining 348 perithecia were sampled from one of the other 11 host species listed in Table 1. All isolates were grouped initially by morphology and a subset was identified with NCBI BLASTn searches using ITS barcoding sequences ( and mountain ash (S. americana) (Fig. 3F), and Cosmospora obscura (Rossman & Samuels) on mountain ash (Fig. 3H).    Fraser fir (Abies fraseri (Pursh) Poir.) ( Fig. 3E and 3C, respectively). Additionally, a novel Corinectria sp. on red spruce (Picea rubens Sarg.) was recovered in this survey across high elevation spruce forests in two states ( Fig. 3D; Fig. 4). In addition to the previously mentioned nectriaceous fungi recovered from perithecia on symptomatic beech trees and nearby co-occurring trees, a member of the Fusarium babinda species complex (FBSC) was commonly isolated from wood tissues of BBD-confirmed American beech trees (31.8% of bark samples), but it was lacking fruiting bodies (Supplemental Fig. 1; Supplemental Table 4). To further investigate this finding, five sites without BBD were identified, and American beech bark tissues were sampled and processed. FBSC was recovered from all confirmed BBD sites but was never recovered from non-BBD sites. Aside from Fusarium babinda, N. faginata and N.
ditissima were occasionally recovered from beech bark tissues collected but only within BBD sites (Supplemental Table 4).

Phylogenetic Analyses
Phylogenetic analyses were performed for single genes (ITS, TEF1, TUB, LSU) and for a four-gene concatenated sequence to infer relationships among Neonectria and Corinectria species recovered in the survey. These analyses were supported with the addition of sequence data from several reference strains for both genera (Table 3).

Representatives of the Fusarium concolor and Fusarium babinda species complexes
were chosen to serve as outgroup taxa that could confirm the identity of Fusarium babinda isolates that were frequently recovered in this study from collected bark tissue samples.
Neonectria neomacrospora was sister to N. ditissima (72%/1.0) and isolates recovered in this survey were sister to selected NCBI reference sequences.

Pathogenicity Assay
A field pathogenicity assay was conducted to further characterize N. magnoliae and test the pathogenicity of N. faginata and N. ditissima isolates originating from American beech on alternative host species including those not previously included in cross pathogenicity assays. Overall, N. ditissima produced significantly larger cankers (p < 0.05) than all other treatments on striped maple, yellow birch, and black birch ( Fig.   5A-C, respectively). N. faginata and N. magnoliae failed to produce cankers significantly larger than the negative control on any of the aforementioned hosts. On American beech, N. ditissima produced larger cankers than N. magnoliae and the negative control, but N. ditissima and N. faginata canker sizes were not significantly different (p = 0.71) (Fig. 5D). N. faginata produced a larger canker than N. magnoliae, although not statistically significant, and negative control on American beech (p = 0.19 and 0.23, respectively). N. magnoliae produced a significantly larger canker on tulip poplar than all other treatments (p < 0.0001) (Fig. 5E). The treatment isolate was recovered from all respective inoculations.

Morphological Characterization of Neonectria magnoliae
The morphological features of Neonectria magnoliae are summarized here and illustrated in Fig. 6. Cankers produced by N. magnoliae on tulip poplar (Fig. 6A) were reminiscent of the perennial target cankers produced by N. ditissima, but these cankers often appeared irregular and less descript, especially for infections on M. fraseri ( Fig.   6B-C). Perithecia occur singly or in aggregates on bark tissue or directly on exposed wood surrounding the outer canker margins and typically emerge from thin stroma tissues as reddish-brown globous body with a distinct ostiole then fade to brown as they age (Fig. 6D). Asci appear truncated and bear eight ascospores (Fig. 6E). Ascospores ((11.3-) 13.1 -15.3 (-17.7) µm x (4.0-) 5.6 -7.4 (-9.1) µm) are uniseptate with roundedends, constricted at the septum, hyaline, and warty on the surface (Fig. 6F-G). N.
magnoliae cultures have a white surface with an orange-red subsurface after 10 d on PDA (Fig. 6H). Microconidial sporodochia and condiophores are regularly produced in culture (Fig. 6I). Sporodochia are slimy masses with a cream to buff color.
Ascospore measurements and general morphological descriptions provided in the type description for N. magnoliae (Lohman and Watson, 1943) were comparable with those observed in this study (Fig. 6). Notable was the infrequency of macroconidia from fresh cultures on general growth media. Interestingly, this trend did not hold for all isolates as several of the N. magnoliae from Fraser magnolia produced abundant macroconidia even in week-old cultures. The only isolate of N. magnoliae from magnolia also showed some sequence divergence from all tulip poplar isolates (Fig. 4). Further characterization of magnolia and tulip-polar isolates are needed to confirm whether these notable differences are biologically significant. The included pathogenicity trial confirmed the pathogenicity of N. magnoliae on tulip poplar but not on other hosts tested (Fig. 5). Interestingly, N. ditissima did not produce cankers significantly different from the negative control on tulip-poplar despite the previous work by Castlebury et al. (2006) confirming tulip-poplar as a host of N. ditissima (Fig. 5) (Table 1). Interestingly, not all hosts had characteristic perennial target cankers (Fig. 2).
Instead, Acer pensylvanicum and Ilex mucronata exhibited vascular cambium necrosis and perithecia production without an apparent outward host response (e.g. callous ridges, bark malformation). Perithecia production also varied across hosts with black birch cankers consistently yielded some perithecia and mountain ash cankers rarely yielding perithecia.
Cross-pathogenicity assays among N. ditissima strains demonstrated the lack of host specificity between isolates recovered from American beech or alternate host species (Fig. 5). Additionally, a recent study confirmed the lack of mating barriers among N. ditissima strains recovered from varying host tree species (Stauder et al., 2020). Together, these results provide additional observational evidence describing the extent of the generalist nature of N. ditissima as a ubiquitous phytopathogen in Eastern North America capable of infecting many hardwood species.
Neonectria faginata was the most recovered member of the Nectriaceae in this study, though it was exclusive to American beech (  (Chapela and Boddy, 1988;Hendry et al., 2002).
Previous assertions of the eventual dominance of N. faginata in the BBD pathosystem were confirmed across our study sites (Houston, 1994;Kasson and Livingston, 2009). In total, N. faginata was recovered from 93.7% of BBD samples collected in this survey (from 13 of 13 BBD sites) while N. ditissima was only recovered from 4.2% of BBD samples (from two of the 13 BBD sites) (Table 1).
Nevertheless, N. ditissima still appears to play a minor role in BBD in its more advanced stages. The pathogenicity trial demonstrated the potentially increased pathogenicity of N. ditissima on American beech when compared to N. faginata (Fig. 5).
While the interaction of the scale insect may have significantly altered the host's physiology and thus, defense responses to fungal infections, these results appear to indicate that an increased virulence of N. faginata may not be responsible for its dominance in the BBD pathosystem.
One additional possibility may be related to its production of perithecia and the resulting inoculum potential. For example, taxonomic descriptions of N. ditissima and N.
faginata illustrate differences in their production of perithecia. In general, N. ditissima is described as bearing fewer, scattered perithecia while N. faginata can be found to produce larger aggregates in higher densities (Castlebury et al., 2006;Lohman and Watson, 1943). Sampling of N. ditissima perithecia across eight hosts generally supports these previous observations (Table 1). Seasonal differences in fruiting could explain differential abundances in these two fungi, but N. ditissima isolates were recovered from beech and non-beech hosts during the same sampling periods in which sampling of BBD trees resulted in high abundances of N. faginata. Therefore, seasonality of fruiting does not appear to have been an apparent factor in this study.
Further, several studies report ascospores as the predominant spore type in the environment and more specifically, the dominant infective spore type (Crane et al., 2009;Lortie and Kuntz, 1963). While these are general descriptive observations and not quantitative comparisons, differences in fruiting habits may explain the eventual abundance of N. faginata due to a significantly increased inoculum potential.
Bionectria ochroleuca was also recovered from BBD impacted beech trees at two total sites in West Virginia and North Carolina, representing 2.1% of all perithecia sampled from beech (Table 1; Fig. 3). This fungus had been previously implicated in BBD although its geographic range was not well studied (Houston et al., 1987). Here we update the known distribution of this fungus to include West Virginia and North Carolina.
Our observations of this fungus on successfully colonized trees supports previous studies, which indicate this fungus is mycoparasitic (Barnett and Lilly, 1962;Jager et al. 1979;Turhan, 1993). Isolates of this fungus were also recovered from American basswood in North Carolina during our surveys (Table 1). This appears to be the first report of this fungus on American basswood in the U.S.
An additional fungus recovered consistently from beech with BBD was a member of the Fusarium babinda species complex (FBSC), although it did not produce perithecia, like all the other Nectriaceae recovered in this study. Instead, members of the FBSC often found colonizing the base of Neonectria perithecia or from necrotic tissue around Neonectria fruiting bodies. After its discovery in single spore plates from Neonectria perithecia, follow-up studies using micro-sampled bark tissue revealed this fungus was present in 31.8% of BBD-positive bark samples at 100% of BBD sampled sites but not present in any beech bark samples taken from healthy asymptomatic trees outside confirmed BBD epicenters (Supplemental Table 4). Its prevalence from BBD impacted stands, but lack of previous reporting, indicates either a localized (albeit unclear) role or novel component in BBD pathosystem. However, a re-assessment of raw data and culture photos taken by the senior author during 2005 -2006 as part of his BBD M.S. research in Maine (Kasson and Livingston, 2009) uncovered a Fusarium with identical morphology to F. babinda in three of three sites sampled for which culture data was collected (Supplemental Fig. 1; See Kasson and Livingston, 2007, Fig. 17A).
This clearly indicates a more widespread association with BBD and one which has not been observed previously despite over 130 years of research on this important pathosystem. One possibility may be that F. babinda is insecticolous, interacting with the scale as populations decline and as Neonectria spp. colonize stems, which is likely given this fungus has been recovered from both gypsy moth and hemlock woolly adelgid populations in the eastern United States (Jacobs-Venter et al., 2018). Another plausible, although not mutually exclusive hypothesis based on the observations of sporulation at the base of viable and nonviable perithecia, is that F. babinda is a facultative hyperparasite of Neonectria. Clearly, more work is needed to test phytopathogenicity and entomopathogenicity, and determine its dominant and facultative lifestyles. Preliminary testing of its ability to degrade tannic acid, a proxy assay for lignin degradation (Kasson et al., 2016) clearly shows that it has strong lignin and cellulose degrading activity.
Finally, in our repeated attempts to uncover non-conventional hosts of N.
faginata, we discovered a putatively novel Corinectria sp. on red spruce in high elevation spruce-fir forests in close proximity to BBD epicenters in VA and WV (Table 1; Fig. 3). Interestingly, the fungus appears to be associated with dead trees but the results of those studies and in-depth investigations into the biology and ecology of this putatively novel species are forthcoming. However, the Corinectria sp. recovered formed a single clade and was divergent from C. fuckeliana and C. tsugae in the concatenated phylogeny (Fig. 4). The third described Corinectria species, C. constricta, was not included in the concatenated phylogeny due to limited availability of sequence data, but based on EF1, isolates recovered in this study were divergent from C.
constricta as well. While this appears to be evidence of a novel species of Corinectria, further investigation is needed.
Taken together, the results of this broad study emphasize the sheer diversity of